singlens

Differentiable single-lens microlensing in JAX: static and annual-parallax light curves, finite sources, limb darkening, and gradient-based inference.

python -m pip install singlens

Requires Python 3.12+ and JAX 0.11.2+. Install singlens[examples] for plotting and inference dependencies. Enable float64 before evaluation.

See the documentation for installation, tutorials, and the API.

import jax
jax.config.update("jax_enable_x64", True)
import jax.numpy as jnp
from singlens import magnification, AnnualParallax

times = jnp.linspace(7980., 8020., 501)
event = dict(t0=8000., u0=0.005, t_e=20., rho=0.01, a1=0.5)
geometry = AnnualParallax(ra_deg=266., dec_deg=-29., t_ref=8000.)
A = magnification(times, **event, parallax=geometry,
                          pi_e_n=0.2, pi_e_e=-0.1)

magnification takes event parameters and returns an array with the shape of times. Omit rho for a point source; supply rho for a uniform disk, or rho and physical a1 for linear limb darkening. Times above are HJD − 2450000 in days. AnnualParallax holds fixed sky geometry; parallax amplitudes remain fitted inputs.

curve = jax.jit(lambda rho: magnification(times, t0=8000., u0=0.005,
                                        t_e=20., rho=rho, parallax=geometry,
                                        pi_e_n=.2, pi_e_e=-.1))
A = curve(0.01)
dA_drho = jax.jacfwd(curve)(0.01)

Use Model, Parameters, and fit_flux for named event parameters, source-position input, and source/blend flux fitting. See the API and examples. Benchmarks compare magnification accuracy, CPU runtime, and inference efficiency with VBMicrolensing and MulensModel.

For source installation and development, see CONTRIBUTING.

Attribution

Adapted from microJAX and the FFT method of Sugiyama (2022). See attribution, NOTICE, and LICENSES for citations and the MIT, Apache-2.0, and BSD-3-Clause license notices.